Mar 22, 2019
This week we’re joined by friend of the pod Dr. Jason Costello, as we take a slight diversion for the usual format and discuss algorithms, Spotify and the future of music.
It may come as a surprise to some people, but algorithmic music has been around in various forms since the 1950s. The first instance of computer-generated composition dates back to 1955-56, where Lejaren Hiller and Leonard Issacson at the University of Illinois used a high-speed computer to create the Illiac Suite. The piece itself was composed by a computer and later performed by a string quartet.
So the concept of computer generated music pre-dates a lot of the music we all know and love. And computers have been used to create, and aid in the creation of music for decades.
Similarly, algorithms have been used for many, many years to do any number of things. After all, an algorithm is simply a set of instructions that are executed in order to solve problems and complete tasks more efficiently.
Algorithms are of course all the rage. They power everything from the computation of an individual’s paycheck by their employers, to the very backbone of the internet itself. A series of computer programs of varying levels and complexities that have become increasingly entwined in our daily lives. In fact, it may prove near impossible for many people live completely algorithm-free lives, such is the way computers regulate large parts of our society.
With increasing advances in artificial intelligence, it should come is no surprise to learn that there many people out there who use algorithms and AI to compose entirely original music. In fact, you can even do yourself by signing up to Jukedeck. Whilst people are still experimenting with how to successfully create music which passes as being made by humans, computers are becoming ever more adept at spitting out music on command.
Spotify and Big Data
The vast majority of people who are reading this right now, almost everyone who listens to this podcast, is signed up to a streaming service. These services are super convenient, and extremely useful for getting you the media you want almost instantly. We won’t bore you by explaining how these work, but you’re no doubt aware of the way these services harvest data about your behaviours on their services. It’s why Spotify can create a daily mix that combines stuff you love with stuff you might not have heard; it’s why they can recommend what to put on a playlist after you add a handful of tracks. They know your preferences.
That data works both ways, and it’s this data which makes up the powerful algorithms these services use to give you what you need before you even know you need it. This data can also be used in reverse to mine the preferences of individuals to create music. Music which may not necessarily be designed for you the individual, but could certainly be created by finding out the traits of the most popular tunes and crafting songs from that data.
It’s not quite Terminator 2 of course, but it does raise many questions.
Jason is on hand field these questions as we get into how and why this is happening, and ponder on what the future holds for royalty free and incidental music. Along the way you will hear some examples of computer generated music, and some artists who have used such tools to create music that they have later went on to augment.
We also chat a lot about how big data effects all the creative industries and so much more.